Jardim et al. (2023): Supplementary Materials

This document contains all data analyses presented in the main article as well as supplementary materials.

The analysis was organized using the make-line pipeline tool provided by targets (Landau 2021). You will find raw data as well as all functions used in the pipeline in the project’s research compendium created with the rcompendium package (Casajus N. 2022).

You can take a look at the pipeline in order to understand some steps (mostly related to tidying data)

 

 

Packages

Code

library(dplyr)
library(tidyr)
library(vegan)
library(ggplot2)
library(targets)
library(ggrepel)
theme_set(theme_light())

Habitat Complexity

Take a look at the structure of the final complexity dataset

Code

tar_load(comp)
knitr::kable(head(comp))

 

First lines of the complexity dataset

Mini	Site	Point	Sample	Branching_density	Broken_density	D_bin	D_gray	L	I	S	Total_Density	Dry_weight	Sphericity	DR1	DR2	DR3
MBE_1	Belle-ile	Belle-ile 1	1	5.8	0.0	1.1231	2.6562	12.85333	9.143333	8.303333	181696	183.61	0.8371338	0.6460062	0.8153846	0.7113589
MBE_1	Belle-ile	Belle-ile 1	2	1.0	0.0	1.0654	2.6463	16.07667	7.883333	5.393333	181696	183.61	0.6122605	0.3354758	0.7669267	0.4903587
MBE_1	Belle-ile	Belle-ile 1	3	3.0	0.2	1.0722	2.6450	15.30000	8.103333	7.190000	181696	183.61	0.7470808	0.4699346	0.8873818	0.5296296
MBE_1	Belle-ile	Belle-ile 1	4	4.8	0.4	1.1078	2.6492	12.91667	8.476667	6.203333	181696	183.61	0.7057078	0.4802581	0.6613704	0.6562581
MBE_1	Belle-ile	Belle-ile 1	5	4.2	0.0	1.1305	2.7019	24.54667	14.523333	8.970000	181696	183.61	0.6088477	0.3654264	0.6434838	0.5916621
MBE_1	Belle-ile	Belle-ile 1	6	3.0	0.4	1.1198	2.6641	16.35667	9.416667	6.783333	181696	183.61	0.6684949	0.4147137	0.7249304	0.5757082

 

Most variables are not normally distributed so we apply a box-cox (Box and Cox 1964) transformation and check correlations between the transformed variables

Code

tar_read(pairscomp)

Figure S 1: ?(caption)

PCA of all observations

Let’s check it without removing highly correlated variables

Code

tar_load(pcatotal)
tar_read(sptot)

Figure S 2: Screeplot of the PCA including all observations and all variables

We had up to 6 relevant axes to look at, but we know we have several highly correlated variables

Code

tar_load(pcatot)
plot(pcatot)

Figure S 3: PCA of all observations

Code

tar_load(pcatotsel)
tar_read(sptotsel)

Figure S 4: Screeplot of the PCA including all observations and selected variables

Let’s check the biplots of the selected variables

Code

tar_load(pcasel)
plot(pcasel)

Figure S 5: PCA of all observations after variable selection

PCA of median values

Code

tar_read(spmedsel)

Figure S 6: Screeplot of the PCA of median values at the point level

Let’s check the biplots of the selected variables

Figure 1

Code

tar_load(pcamed)
plot(pcamed)

Figure 1 (main article)

Check correlation between the centroids of the total PCA and the scores of the reduced PCA

Code

tar_load(corcomp)
plot(corcomp[[1]])

`geom_smooth()` using formula = 'y ~ x'

Code

plot(corcomp[[2]])

`geom_smooth()` using formula = 'y ~ x'

Code

tar_read(cortestcomp1)
tar_read(cortestcomp2)

    Pearson's product-moment correlation

data:  med_sel2$PC1_c and med_sel2$PC1_score
t = -12.623, df = 28, p-value = 4.471e-13
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
 -0.9626648 -0.8416107
sample estimates:
       cor 
-0.9222483 

    Pearson's product-moment correlation

data:  med_sel2$PC2_c and med_sel2$PC2_score
t = -14.294, df = 28, p-value = 2.165e-14
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
 -0.9702598 -0.8722282
sample estimates:
       cor 
-0.9378045 

(a) PC1

(b) PC2

Figure S 7: Relation between Full PCA centroids and Reduced PCA scores

Fauna

Code

tar_load(faunadata)
knitr::kable(head(faunadata))

 

First lines of the original macrofaunal data

Point	Site	Replicate	Method	Habitat	Year	Season	Date	Species	Abundance
Belle-ile 1	Belle-ile	1	Grab	Maerl	2007	Spring	2007-03-11 00:00:00	Abludomelita gladiosa	1
Belle-ile 1	Belle-ile	1	Grab	Maerl	2007	Spring	2007-03-11 00:00:00	Abra alba	1
Belle-ile 1	Belle-ile	1	Grab	Maerl	2007	Spring	2007-03-11 00:00:00	Ampharete spp.	1
Belle-ile 1	Belle-ile	1	Grab	Maerl	2007	Spring	2007-03-11 00:00:00	Amphipholis squamata	2
Belle-ile 1	Belle-ile	1	Grab	Maerl	2007	Spring	2007-03-11 00:00:00	Anapagurus hyndmanni	1
Belle-ile 1	Belle-ile	1	Grab	Maerl	2007	Spring	2007-03-11 00:00:00	Aonides oxycephala	3

 

Data availability

Code

samplespoint <- faunadata %>% 
  mutate(Date=as.factor(format.Date(faunadata$Date,"%Y"))) %>% 
  distinct(Site, Year, Point, Replicate) %>%
  group_by(Site, Year, Point) %>%
  mutate(Year = as.factor(Year)) %>% 
  summarise(count = n()) %>%
  ungroup() %>%
  mutate(State = factor(case_when(count==3 ~ "Good", count < 3 ~ "Sample(s) missing", count > 3 ~ "Too many samples")))

ggplot(samplespoint, aes(x=Year,y=Point,fill=State)) +
  geom_tile(alpha=0.5) +
  scale_fill_manual(values=c("palegreen3","lightblue", "indianred")) +
  geom_text(aes(label=count)) +
  theme(text=element_text(size =15), axis.text.x = element_text(angle = 90, size = 15), legend.position = "bottom")

Figure S 8: Number of grab samples taken every year at each point

Empty cells show points where no samples were collected at a given date due to field constraints.

Alpha diversity

Total richness

Code

# Total richness
tar_load(faunaclass)
total_rich <- nrow(faunaclass %>% 
  distinct(., Species)) #725 species

# Total richness by sediment position

epi_rich <- nrow(faunaclass %>% 
  filter(Position == "Epifauna") %>% 
  distinct(., Species)) #341 epifaunal species


inf_rich <- nrow(faunaclass %>% 
  filter(Position == "Infauna") %>% 
  distinct(., Species)) #266  infaunal species


int_rich <- nrow(faunaclass %>% 
  filter(Position == "Interstice") %>% 
  distinct(., Species)) #118 interstitial species

There are 725 species in total, 341 being epifaunal, 266 being infaunal, and 118 interstitial.

Relative abundance

Figure S2

Code

tar_read(relab_phy)
tar_read(relab_ord) 

(a) Main Phylla

(b) Main Orders

Figure S 9: Average Relative Abundance of Main Phyla and Orders

p.s.: it is normal to have NAs for some orders as higher classification for some of the species is ambiguous or unknown.

Code

tar_read(relab_arth)
tar_read(relab_ann) 
tar_read(relab_mol)

(a) Arthropoda

(b) Annelida

 

(c) Mollusca

 

Figure S 10: Families Average Relative Abundance for Each Main Phyla

References

Box, G. E. P., and D. R. Cox. 1964. “An Analysis of Transformations.” Journal of the Royal Statistical Society. Series B (Methodological) 26 (2): 211–52. https://www.jstor.org/stable/2984418.

Casajus N. 2022. Rcompendium: An r Package to Create a Package or Research Compendium Structure. https://github.com/FRBCesab/rcompendium.

Landau, William Michael. 2021. “The Targets r Package: A Dynamic Make-Like Function-Oriented Pipeline Toolkit for Reproducibility and High-Performance Computing.” Journal of Open Source Software 6 (57): 2959. https://doi.org/10.21105/joss.02959.